Tobacco Documents Online

( Sept. 29, 1982 MEMORANDUM TO : W. Ty HOYT FROM: Hilda and Leonard Zahn SUBJECT: International Cancer Congress Seattle, Sept. 8-1i5, 1982 It was a big and busy meeting -- 4,000+ papers and about 7,000-8,000 visitors -- and it was difficult for many to decide which of the dbzens of consecutive sessions to attend. On the other handi, the nature of the meeting was such that one did not expect any major scientific advances to be reported; the thrust, especially in the plenary sessions, was toward reviews by I experts. While there was considerable interest in presentations dealing with oncogenes, perhaps the hottest area in.cancer... research today, the several thousand individual papers and', posters were looked at solely as offering bits and pieces of information. As could be expected, a comparative few were excellent, most were run-of-the-mill, and some were quite bad. -(It should be noted that none of the offerings really underwent peer review before acceptance.) factors. DeVITA, director of the National Cancer Institute, held a press conference and'when it was done, one had difficulty trying to, recall if he'd'really said anything substantive. One point that can be recalled: the government erred several years ago in assigning suchia large percentage of cancer cases to.occupational . The usual panoply of cancer industry leaders and politicians, especially from the U.S., was present, publicly propagandizing for more funds and greater cooperation for the solutions that seem to lie just around the corner. VINCENT The antismoking movement was there in full array -- from A to Z -- AUERBACH to ZELEN, Marvin (who last year, at a smoking control workshop in Japani, recommended that cigarette packages have coupons -- RADS -- to be redeemed for treatment for lung cancer patients). Most of themisaid what was familiar and ~ expected, but a few had some comments that were a bit surprising (see especially the items below on RADFORD and HIRAYAMA). V There were several interesting reports by Japanese scientists that stressed the need to look beyond cigarette -smoking inithe genesis of lung cancer. A number of sessions dealt withismoking cessation role of vol'untary societies in controlling cancer. eonard a~" ~ __ PUBLIC RELATIpNSCOUNSEL anf tdA~ssocialesInG O GJ j and the G.7 These, for the ~ ~ N (P. O. BOX 223) 13 LINCOLN ROAD • GREAT NECK, N.Y. 11021 •(212) 895-7445

2. most part, were more of the same old, tired stuff'and seemed to have been scheduled simply to provide a reason for many foreigners to attend as speakers or "learners." .~ The press relations man for the American Cancer Society and his counterpart at NCI were able to schedule people of their choice for press conferences. The ACS man thus brought in GARFINKEL to discuss low-tar, low-nicotine cigarettes. This particuir session generated national and, it is assumed, international press attentioni. Press people were present from many countries, but the U.S. contingent was somewhat thin. No newspaper reporters came from such major cities as New York, Washingtoni, Chicago, or San Francisco. It didn't matter, though, because the two major newswire services were represented. The usual press complaint was frequently heard -- there's not.muich here to write about. It was learned during the meeting from a Japanese acquaintance that MITSUiO1SEGI, well-known for his periodic collections of international cancer statistics, died last May. Wee were saddened to learn of the death on Sept. 12 of JOSEPHI BERKSON. V/ Highlights of the meeting: 1."Chronological analysis on the relation between lung, cancer and cigarette smoking".-- WATARU MORI, Tokyo, and RYOJ!Ii SAKAI, Naha, Japan. The upshiot of this study reported by Mori (he's apparently a pathologist, Sakai an epidemiologist), who gave the paper, is that there seems to be more than smoking involved, in the development of lung cancer. Their data show a greater increase in lung cancer among nonsmokers than among smokers. Mori said at the outset that the incidence of lung cancer among autopsies (apparently at his institute)' has increased from 4.6& to 8.8% in the last 20 years and that lung cancer is now the second most common malignancy in his country. The study covered 6,610 autopsies of adult lung cancer cases (4,269 males, 2',341 females) with a history of'smoking that were performed at the University of Tokyo. The overall time period was 193'6-55 and 1959-78. The cases were divided into four groups according to year of birth,: 1936-45, 1946-55, 1959-68, 1969-78. Daily smo'king amounts were broken into four categories: none, 1-19, 20-39, and 40+. Conclusions: A. Lung cancer incidence was significantly higher in smokers of bothisexes in all groups.

C 3. B. The incidence of lung cancer increased from 1.9% to 8.3% during the period. Inismokers the increase was from 4.4% to 11.8%, but ininonsmokers it went from 1.0% to 5%. I C. The lung cancer incidence in nonsmokers' autopsies of the 1969-78 period was 5.6% for males and 4.6% for females, values'as high as were found in smokers inithe 1936-45 and 1946-55 periods. D. The relative risk of developing lung,cancer seems to have decreased during the period and so does the dose-response relationship between smoking and'lung cancer. It seems clear that other factors have become more important than cigarette smoking in the causation of lung cancer. . ERNST WYNDER, who was cochairing the session ("Tobacco '. Chewi'ng and Smoking,in Cancer Causation"), commented that the smoking histories of the cases did not seem to be very precise. Mori said they came from death certificates and1while they may have been "rough," there was no doubt about the accuracy of the diagnoses. 1 , ~ 2. "Time trend data of'histological subtypes of lung cancer in Japan: observations at the Cancer Institute hospital fromi 1961-78"' -- this was a poster presentation with E. TSUCHIYA as lead author. The lung cancer mortality rate in Japan has increased ' steeply in both sexes: in 1978, as compared to 1947, it was 11 times higher in men and 9 times higher in women. To verify the presumption that the increased number of cases contained more squamous cell carcinomas than other cell types, a study was done in which histologic slides were reviewed andi chronologic changes in histologic subtypes were examined. The study material consisted of 640 resected and nonresectedilung cancer cases the Cancer Institute Hospital. The number of lung cancer cases increased 5 times in malles and 8 times in females in the 1961-78 period. In males, adenocarcinomas decreased slightly until 1973 and then began to increase. Squamous cell carcinoma had increased slightly until 1969 and remained steady thereafter. In females, adenocarcinomas had increased and squamous cell cancers decreased from 1972 on. All patients except the adEnocarcinomas were smokers. However, 18% of the adenocarcinoma patients were nonsmokers. This figure is simila'r to that for the general population for the 19174-81 period. Case distribution by histologic subtype was as follows: adenocarcinoma 40%; squamous cell 28%; small cell 19%; large cell 8%. The ratio of adenocarcinoma to squamous cell was 1.0 in males and 4.11 in females. The ratio of each subtype did not change distinctly during the period, but since 1973-74, adenocarcinoma has increased slightly in both sexes. Squamous, small andilarge

4. C cell cancers are related to smoking but adenocarcinoma has-no, or weak (if any), relationship~to smoking. Intrinsic carcinogenic factors must be considered in the genesis of adenoca:rcinoma of the lung. -U V 3. "Trends of lung cancer incidence and their histological distribution in.Osaka" -- AYA HANAI, Osaka, Japan,. Lung cancer mortality and morbidity have been increasing in Osaka, a city of 8.5 million persons engaged mostly in commerce and industry. The disease is now the second most frequent site of cancer in men though it is still much less common in Japan than iniOccidental countries. Of the 13,037 new lung cancer cases registered between 1967-78, only about a third were histologically verified...(A person in the audience suggested tha:t this created a bias.) However, H'ianai said her group is now confirming 65% of the.cases and the results seem to hold. In males, epidermoid cancer (E) was found to be the most prevalent type (46%) with adenocarcinoma:(A), which was more common in younger men, next at 34%. Females had more A, followed by E and undifferentiatedl(U). In both sexes, A and U are increasing in Osaka. . , Though E is most prevalent among older men, its increased incidence recently has been only 6%. Among femaLes, who usually are nonsmokers, U has increased 94%. Therefore, the association of smoking with the increased risk of E and U types of lung cancer needs to be reexamined. f 4. "Determination of smoking-specific lung cancer rates in, epidemiologic studies",,-- EDWARD RADFORD, Pittsburgh. This was a somewhat convoluted and often less-than-clear paper in which Radford purported to show the additive effect of smoking and radon daughter exposure in causing lung cancer, at least in' Swedish miners he studied. Along the way he found a higher relative risk for nonsmokers than for smokers, but he had an explanation for this. Radford~ began by noting that while lung cancer clearly is a problem, there's a growing indication that "there are other causes of lung cancer besides smoking and one of the important probl!ems that we fin~d in epidemiologic research on lung cancer is clearly thalt smoking is such a variable...the time for qualitative assessments of the confounding effects of smoking is over...I'thi,nk we have to start dealing with this quantitatively." In order to get smoking-specific disease rates (disease in general, not just lung cancer), one must have the smoking status for both the study and reference population~s. Also necessary are sex, age-specific d~ata for history of smoking over an appropriate time periodi(which must be determined), pack-years, and

C 5. C cumulative dose. All are "tricky" issues. One also:must know the fraction of current smokers by age with attentionito former smokers and~type of smoking, the intensity of smoking by type, and the relative risk of the disease by smoking category compared to nonsmokers in the population. It's not easy to get relative risk data: fromithe various national population studies, nor is it easy to get bothi year-specific and age-specific data that may be needed. That's why the whole subjiect is solcomplex; he therefore made a few "assumptions. " He surveyed the age-specific relative:risk for smokers comparedito nonsmokers. Using his own formula, he studied 556 active and retired Swedish miners, determining those who smoked and those who did not as a function.of age at interview. (He combined exsmokers and never-smokers.) He plotted never, long-term ex-, recent ex- and current smokers at each range, and constructed an age-specific formula to show th~e fraction of smokers. The study included data fromithe 1963-72 Swedish national .smoking survey (R'UNE CEDERLOF'et al). The miners were found to have a: somewhat higher rate of giving up smoking than did the national population. The Swedish national data sh~ow that pipe smokers have almost the same relative risk for lung,cancer as do cigarette smokers ("rather to everyone's surprise"). So pipe and cigarette smokers were pooled in the miners' study. The Swedishistudy showed a relative risk of 7 (for current, occasional and exsmokers), and the miners' study showed a relative risk of 7.4. "The miners were slighitly heavier smokers, but there was a smaller fraction of them as a function of age because they gave up smoking at a more rapid, rate." In some cases, heart disease, e.g., on~e would not necessarily assume the relative risk would be independent of age, but in the case of lung cancer "we made that assumption." . All the information was put together to obtainiage-specific rates that were applied to 1,274 miners (it was unclear whether the 5561miniers were part of this group) alive in,1951 and followed through 1976. In ord~er to get a; quantitative picture of the relative importance of smoking inireliation to radon daughter exposure,' the data were analyzed for expected lung cancer cases, with and without correction for smoking,. At the end~, it wa:s see:n that the relative risk for nonsmoking miners was "substantially higher" than that for the smoking miners. The absolute risks were not very different. Ra:dford'conicluded: "This is almost diametrically opposed to practically everything you may have heard elsewhere in this

6. meeting whenever this relationship of cigarette sd'moking to pulmonary carcinogenesis has been presented. I submit that this is probably k-more typical result, namely the additive effect of smoking and the carcinogen, in this case alpha radiation from radon daughters." ./ 5. "Smoking and chewing in the developing world: the epidemiological approach" -- TAKESHI HIRAYAMA, Tokyo. In Japan,, considered a developing country, cigarette smoking is still on the rise and is now at the highest level in history. Lung cancer is the Number I killer in both sexes (he must have meant in smokers, but he didni't say so). Seventy-five percent of men aged 20-39 years and 70°/'0 of those over 401years smoked in 1981. Japanese women also are smoking more because of "intensive" advertising and promotionifocussed onithem, especially young women. He has found that women are evenimore susceptible than men to smoking-related diseases for the same amount of cigarettes smoked. Long lists of these diseases and disorders included: cerebral vascular, ischemic and other heart diseases, asthma, eczema, and cancers of such organs as the buccal cavity, pharynx, larynx, esophagus, stomach, and liver. Cancers of the gallbladder, cervix and ovary are very closedly related to smoking. Breast cancer in women and prostate cancer in men stilll seem to be exempt from the effects of smoking. - - The age at which a woman starts smoking is very ;mportant. The earlier she starts, the higher the risk of these cancers, independent of the total number of cigarettes every smoked. It appears that age is related to the initial stage of cancer while the number of cigarettes is related' to promotion. Briefly discussins his passive smoking research, he said the nonsmoking womenlin his study were married to men who smoked 50., or more cigarettes daily (is this the first time he's disclosed that figure?). In addition to lung cancer, nonsmoking wives of smokiing husbands also have higher risks for stomach and cervical cancer, ischemic heart disease, etc. It seems passive smoking is affecting the incidence of chronic diseases as well as lung cancer. Whili'e passive smoking data from the American Cancer Society study (Garfinkel) aren't significant, their trend is similar to data from his study! r U3'73494s Hirayama also gave statistics on cigarette smoking from some large cities in Asia and discussed the problems of betel chewing and bidi (local cigar) smoking, especially in Southeast Asia.

(Hirayama: spent several hours hanging around the press room one day until he was finally interviewed by a reporter for a local newspaRer, at the request of th~e ACS press relations director. A story the next day quoted!Hirayama on his passive smokiing work. He was quoted that results similar to his alsolwere found in studies done in Greece and by the ACS. I 6. Two adjoining posters were presented by CAROL HENRY on the CTR-sponisored project at Microbiological Associates, Inc., Bethesda, Md~. Th~e first, with grantees RICHAR',D RASMUSSEN and, WILLIAM BENEDICT also listed, was on "The effect of exposure tp whole cigarette smoke on short-term endpoints in BC3F1/Cum mice" and the second dealt with lifetime exposure of the mice to 2R'1 cigarettes. Findings of the first poster: A. Exposure of the mice to 3A1 (high tar, high nicotine) cigarettes resulted in about a 3-4-fold increase in aryl hydrocarbon hyd'roxylase (AHH) activity after one day's exposure or after repeated~daily exposures over 39 weeks. Renal AHH activity was similarly increased in mice chronically exposed to smoke. B. Ornithine diecarboxylase:activity in lung: tissue was increased about 2-fold in mice exposed to smoke daily after three or six months. C. Replicative DN!A synthesis in lung tissue was increased 2-5-fold in exposed~mice after three or 58' weeks. D. Prolonged daily exposure greater than 13 weeks to smoke resulted in increased lung/body weight ratios an~d inicreased'n protein and! hydroxyproline content of lung tissue. E. Exposure for as short a period as one week andias long as 46 weeks resulted in a 2-fold increase in the frequency of sister chromatid exchanges. This increase persisted even,after a one-week cessation of exposure. F. There was an increase in DNA synthesis rates after nine weeks' exposure. Exposure to 3A1, but not 2R1, cigarettes resulted in a 50% decrease in lung DN~A repair. Thus, the poster concluded~, cigarette smoke is capablie of causing specific and quantifiabl!e biologic effects in a murine model system. Findings of the second poster: A. Daily exposure for 110 weeks resulted in a time to 50% survival of 110!, 1019 and 106 weeks for smoke, sham and shelS control groups, respectively. B'. Body weights of both smoke- and sham-exposed animals were not different during the 156 weeks of observation. Shelf control animals had significantly higher body weights than smoke- or sham-expdsed animals. C. Over the 110-week exposure period,about 0.8 grams of tobacco particulate matter (TPM) were generated, resulting in about 625:microgr,ams/TPM deposited in the respiratory tract per mouse per week. D. The mean fl373,4s4*7 carboxyhemoglobin l~evels per day in the smoke-exposed animals were 17.2% compared to 1.4% for sham-exposed and 1.8% for shelf controls. E. Accumulation of pigmented alveolar macrophages was first seen after 48'weeks' dail'.y exposure. The incidence in the last 16 weeks of exposure was 39% and this fell to about 20% in

8. the final 40 weeks. F. Inianimals randomly sacrificed during the 1!10 weeks of exposure there were 7 malignant lung,tumors in the smoke-exposecL.mice compared to none in the sham-exposed. G..In animals that died or were sacrificed~ when moribund in the entire observation period of "1561weeks, 16 mice exposed to smoke died of lung cancer compared to 7 sham-exposed. An incidental observation of fibrosarcomas (FS) of the head and'neck occurred after 72'weeks on test. These tumors were mnetastatic and led to the death of the animals. Significantly more animals died from the FS in the smoke-exposed groups compared to the sham-exposed )by 116 weeks on test). There were no significant differences between the smoke- and sham-exposed groups for FS of other tissues and organs. (Henry told visitors the experiments indicate that tobacco smoke'has, at best, some weak carcinogenic activity. A paper on the second poster is being readied for submissionito the journal "Cancer Research," but Henry is unsure whether it will be accepted because the results generally show a low level of significance. A paper on the short-term study is to be published in "Carcinogenesis." V 7. "Smoking,and cervical cancer: evidence from a study of cervical dysplasia" -- E.A. CLARKE, Toronto. Clarke said at the outset that she believed cigarette smoke is a squamous cell carcinogen and that the study she was reporting provides further support for the smoking-cervical cancer claim first proposed in 1977 (by W. WIN!KELST'EIN)'. - Her study was done on cervical dysplasia, which could be a precursor of cervical cancer.;It was a case-control study covering 2501patients and 500 controls aged 25-34 years. In contrast to the controls, the cases were of younger age at first iintercouirse andi at first pregnancy, had a greater number of sexual partners, lower family incomes, and fewer of them had finished high school. More of the cases than controls were smokers, leading to a relative risk of 4.85 for developing cervical dysplasia for smokers compared'to 2.51 for nonsmokers. The trend~was for an increased risk in the younger age groups, a finding supported by recent data from a Utah study. The risk of current smoking persisted after adjustment for the number of sexual! partners and age at first intercourse. There was a 4% increase per year in risk per pack year smoked and a 12% increase per number of sexual partners. Therefore, the evidence is there for the relationship between smoking,and cervical dysplasia. 03734948 Some questions remain: What is the mechanism(s)? Is the ingredient (unidlentifiedl in tobacco smoke a promoter or a cocarcinogen that works with another carcinogen associated with

sexual behavior?'Is there a tobacco nitrosamine that is organ~ specific?' In later discussion, WYNDER'commented this was an example of what he called "the epidemiology of weak associations." There znay be certain nutritional deficiencies in female smokers that may effect cervical d~ysplasia. Clarke!said she hasn't yet analyzed data on alcohol intake an~d oral contraceptive use in her population. 8. "The changing cigarette: its influence on mortality rates" -- LAWRENCE GARFINKEL, AmericaniCancer Society Vice President of Epidemiology and Statistics, New York. He presented~ data collected between 1955-72 fromithose enrolled inithe ACS' prospective study. (Much of what he discussed', in Sea:ttle was reported last spring at the annual ACS science writers' seminar. Also,•it seems it was reported and published several years before that by E. CUYLER HAMMOND, his predecessor at ACS.) At both a press conference and~fromithe platform, Garfinkel, who also has the title of Director of Cancer Prevention of ACS, said findings indicated a"d'efinite' benefit" for smokers of low T/N cigarettes compared to those who smoke high T/Nbrands. The difference is more pronounced for women thanifor men, but smokers of low T/N cigarettes were found: to have a 26% lower lung cancer mortality thanidid smokers of high T/N cigarettes. On the passive smoking,issue: The data show that for nonsmoking women whose husbands smoke less than 20 cigarettes daily and for those whose husbands smoke more than 2&daily, the observed vs. expected dethirates were not significant (at the. 0.05 level, and were 1.27 and 1.10, respectively). The differences were found to be nonsignificant inimatched-group analysis. Adjusted lung cancer death rates were 1.37 and 1.04 for women whose husbands smoked~less than 20 cigaretters/day or 20+/day, respectively, in comparison to women with.nonsmoking husbands. Later on (at the press conference), he said the effects of passive smoking really can't be determined from either the Japanese (HIR'AYAMA) study or that by the ACS. The new ACS study, which started'.September 1, should give more definitive data in time. Garfinkel also made these points: -- There's only a slight difference in coronary heart disease mortality between low and highiT/N' smokers. _03,734g49 -- There's an advantage for those who continue tosmoke to use low T/N cigarettes. It's easier for them t&give up smoking (a range of 4-7%). They also seem to smoke fewer cigarettes daily.

10. -- While some studies have claimed that low TIN smokers smoke more and inhale more deeply, his data show that the majority of low TIN smokers smoke at about the same rate they did ~ when they used high T/N cigarettes. In the short run, when people take up low T/N cigarettes they tend to smoke more, but they reduce their consumption over the long term. -- No drop in lung cancer is yet apparent. However, the lung cancer mortality rate should begin to fall (especially in men in 5-6 years) because many smokers have quit or have switched to low T/N brands and because younger people who smoke are using low T/N cigarettes. -- The picture is different for women because they began smoking later. Lung cancer mortality is low in women aged 35-44; women under 45 either decided, early in life not to smoke or took up low T/N cigarettes. -- The number of cigarettes smoked daily is far moree important than the T/N content. -- He has no direct evidence of the harmfulness of additives in low T/N cigarettes, but he thinks it important for the tobacco industry to divulge the nature of everything put into these cigarettes. -- The ACS, with OSCAR AUERBACH, is conducting,a study in four hospitals in which lung,cancer victims will be traced from records; interviews will be done with family members to determine smoking histories. Auerbach will review all available slides of lung tissue. (Garfinkel's data covered two periods: 1960-65 and 1966-72'. Diseases considered in his analysis included cancers of the lung,, buccal cavity/pharynx, esophagus, larynx, bladder, and'pancreas, and nonmalignant conditions such as coronary heart disease, stroke, aortic aneurysms, emphysema, liver cirrhosis, and stomach ulcers.) 8. "Positive correlation between high AHH activity and primary lung cancer in humans" -- a cooperative CTR'project involving RICHARD K0'URI of Microbiological Associates and T. McLEMORE of Houston. Blood samples from 21 lung cancer (20 of them smokers)' and 30 noncanceer patients at a Houston VA hospital were sent by McLemore to Kouri, who used a new technique to determine' AHIH levels. The new technique employs an improved, lymphocyte AHH culture system that reportedly solves the problems of previous AHH studies that had poor reproducibility of the culture system. 03'734954 A "striking" association was found between the presence of lung cancer and highiAHH levels. The 14 highest AHH activity levels were found inithe lung cancer group. Twenty-one of the noncancer group had low AHH activity. A simillar 1iung canicer-high